JOURNAL OF MEDICAL INTERNET RESEARCH Taneja et al

Original Paper Social Media and Research Publication Activity During Early Stages of the COVID-19 Pandemic: Longitudinal Trend Analysis

Sonia L Taneja1*, MD; Monica Passi1*, MD; Sumona Bhattacharya1, MD; Samuel A Schueler1, MD; Sandeep Gurram2, MD; Christopher Koh3, MHSc, MD 1National Institutes of Diabetes and Digestive and Kidney Diseases, Digestive Disease Branch, Bethesda, MD, United States 2Urologic Oncology Branch, National Institute, National Institutes of Health, Bethesda, MD, United States 3Liver Diseases Branch, National Institutes of Diabetes and Digestive and Kidney Diseases, Bethesda, MD, United States *these authors contributed equally

Corresponding Author: Christopher Koh, MHSc, MD Liver Diseases Branch, National Institutes of Diabetes and Digestive and Kidney Diseases 10 Center Drive Clinical Research Center, 5-2740 Bethesda, MD, 20892 United States Phone: 1 301 443 9402 Email: [email protected]

Abstract

Background: The COVID-19 pandemic has highlighted the importance of rapid dissemination of scientific and medical discoveries. Current platforms available for the distribution of scientific and clinical research data and information include repositories and traditional peer-reviewed journals. In recent times, social media has emerged as a helpful platform to share scientific and medical discoveries. Objective: This study aimed to comparatively analyze activity on social media (specifically, Twitter) and that related to publications in the form of preprint and peer-reviewed journal articles in the context of COVID-19 and gastroenterology during the early stages of the COVID-19 pandemic. Methods: COVID-19±related data from Twitter (tweets and user data) and articles published in preprint servers (bioRxiv and medRxiv) as well as in the PubMed were collected and analyzed during the first 6 months of the pandemic, from December 2019 through May 2020. Global and regional geographic and gastrointestinal organ±specific social media trends were compared to preprint and publication activity. Any relationship between Twitter activity and preprint articles published and that between Twitter activity and PubMed articles published overall, by organ system, and by geographic location were identified using Spearman's -order correlation. Results: Over the 6-month period, 73,079 tweets from 44,609 users, 7164 journal publications, and 4702 preprint publications were retrieved. Twitter activity (ie, number of tweets) peaked in March 2020, whereas preprint and publication activity (ie, number of articles published) peaked in April 2020. Overall, strong correlations were identified between trends in Twitter activity and preprint and publication activity (P<.001 for both). COVID-19 data across the three platforms mainly concentrated on pulmonology or critical care, but when analyzing the field of gastroenterology specifically, most tweets pertained to pancreatology, most publications focused on hepatology, and most covered hepatology and luminal gastroenterology. Furthermore, there were significant positive associations between trends in Twitter and publication activity for all gastroenterology topics (luminal gastroenterology: P=.009; hepatology and inflammatory bowel disease: P=.006; gastrointestinal : P=.007), except pancreatology (P=.20), suggesting that Twitter activity did not correlate with publication activity for this topic. Finally, Twitter activity was the highest in the United States (7331 tweets), whereas PubMed activity was the highest in China (1768 publications). Conclusions: The COVID-19 pandemic has highlighted the potential of social media as a vehicle for disseminating scientific information during a public health crisis. Sharing and spreading information on COVID-19 in a timely manner during the pandemic has been paramount; this was achieved at a much faster pace on social media, particularly on Twitter. Future investigation could demonstrate how social media can be used to augment and promote scholarly activity, especially as the world begins to increasingly

https://www.jmir.org/2021/6/e26956 J Med Internet Res 2021 | vol. 23 | iss. 6 | e26956 | p. 1 (page number not for purposes) XSL·FO RenderX JOURNAL OF MEDICAL INTERNET RESEARCH Taneja et al

rely on digital or virtual platforms. Scientists and clinicians should consider the use of social media in augmenting public awareness regarding their scholarly pursuits.

(J Med Internet Res 2021;23(6):e26956) doi: 10.2196/26956

KEYWORDS coronavirus; COVID-19; social media; gastroenterology; SARS-CoV-2; research; literature; dissemination; Twitter; preprint

utilized social media (eg, Twitter, Facebook, TikTok) for Introduction communication and to facilitate interdisciplinary discussion COVID-19, caused by the novel coronavirus SARS-CoV-2 [13,14]. (severe acute respiratory syndrome coronavirus), emerged into Social media platforms such as Twitter are social networking public view in December 2019 and resulted in a pandemic that services through which any user or organization with an account, has affected six continents, and it continues to indiscriminately including those belonging to the scientific and medical affect individuals of all ages, races, and ethnicities. According communities, can share information and achievements. to the World Health Organization, there have been more than Compared to other professions, the health care community has 33,000,000 confirmed cases of COVID-19 globally, including been relatively reluctant in utilizing social media for professional more than 1,000,000 deaths reported by the end of September purposes related to concerns on its potential impact on 2020Ðonly 9 months after its emergence [1]. With time, employment, medicolegal liability, and relationships among different countries faced surges in cases straining their health patients and colleagues [15-17]. Nonetheless, as these platforms care systems in unprecedented ways. It is during these times continue to gain global acceptance and utilization, the ability that the rapid dissemination of information related to this highly to collect and analyze data from social media platforms has contagious virus and its management has been crucial. become essential in understanding health care±related needs, Even though initial experiences related to COVID-19 primarily shifting public health interests, and highlighting areas for further described respiratory complications, reports of medical study [18]. In this study, we aimed to explore gastrointestinal/gastroenterology (GI) involvement became COVID-19±related social media activity pertaining to the fields more evident with increased clinical experience [2]. Although of gastroenterology and hepatology during the initial 6-month the extent of GI involvement with COVID-19 was uncertain period of the COVID-19 pandemic, when knowledge about the based on early published experiences, it was postulated that this virus was new and limited. Furthermore, we aimed to compare could be substantial due to the identification of the entry activity on social mediaÐspecifically TwitterÐwith that via mechanism of SARS-CoV-2 that utilizes the -2 more traditional channels of medical information sharing and (ACE2) receptor pathway, which is found throughout the GI distribution such as publications in medical journals and preprint tract, liver, and pancreas. Given the pathogen's similarities to repositories. the coronavirus known to cause severe acute respiratory syndrome (SARS) and Middle East Respiratory Syndrome Methods (MERS), investigators suspected that prior experiences with these preceding could provide insight into the current Data Collection for COVID-19±Related Twitter COVID-19 pandemic. Thus, GI luminal manifestations, the Activity involvement of the liver and pancreas, and the management of Data were collected using the publicly available Twitter unique GI patient populations were all considered areas of analytics platform Symplur Signals [19]Ða health care social clinical and research interest [3-5]. Considering the rapid spread media analytics platform that utilizes algorithms with natural of COVID-19 and the consequent interruption of health care language processing to provide in-depth information on Twitter services across multiple fronts, the publishing of international activity. Data on topics related to COVID-19 were collected by experience with COVID-19 along with frequent updates in performing specific searches categorizing the topics by organ clinical guidance documents have assisted the GI community system (Multimedia Appendix 1). Data were captured over the in managing this novel disease [6-9]. Furthermore, in an effort first 6 months of the COVID-19 pandemic, from December 1, to mitigate the spread of infection, endoscopists have 2019, to May 31, 2020. In an effort to capture the longitudinal encountered significant changes to endoscopic practices by evolution of social media use during this period, each month adopting new preprocedure regulations, use of enhanced was split into half (ie, day 1-15 and day 16 to the end of the personal protective equipment, and the rearrangement of month). Data collected included total number of tweets and endoscopy units to facilitate social distancing [10,11]. retweets, total number of impressions, total number of users, Additionally, with the implementation of national ªlockdowns,º and user data, including the place of originÐby country the ability to share clinical experiences, analyze medical data, (globally) and by state (within the United States). The ratios of and disseminate management strategies for COVID-19 has tweets per Twitter user and impressions per tweet were also become reliant on electronic media [12]. During these calculated (Figure 1). Definitions of these terms can be found unprecedented times, the medical community has increasingly in Multimedia Appendix 1.

https://www.jmir.org/2021/6/e26956 J Med Internet Res 2021 | vol. 23 | iss. 6 | e26956 | p. 2 (page number not for citation purposes) XSL·FO RenderX JOURNAL OF MEDICAL INTERNET RESEARCH Taneja et al

Figure 1. Study flow diagram outlining the data extraction method from 3 media platforms analyzed (Twitter, PubMed-NCBI, and medRxiv or bioRxiv). *Each month of the specific study period was split into half-month intervals for the purposes of analysis. **Duplicate publications from separate searches were individually reviewed and reorganized into the most appropriate subject group in order to eliminate potential for publications to be accounted for more than once. ***A follow-up review of preprint articles pertaining to COVID-19 and gastroenterology that ultimately resulted in formal peer-reviewed journal publications was performed for July 2020. IBD: inflammatory bowel disease; GI: gastrointestinal.

specific search terms used are detailed in Multimedia Appendix Data Collection for COVID-19±Related Preprints and 1. All resulting from PubMed searches were recorded, Publications and the search results were further filtered by half-month time Preprint articles are research manuscripts shared publicly before intervals, identical to the search methods used for Twitter , which allows for rapid dissemination of content and preprint articles for the purposes of comparison. information, thereby helping to inform policy and clinical For both preprints and publications, articles were further practice in a timely manner. Preprint repositories have gained subgrouped by organ system topic. Duplicate publications from considerable attention over the course of the COVID-19 separate searches were individually reviewed and recategorized pandemic and have been increasingly utilized for the into the most appropriate subject group, thereby eliminating the dissemination of crucial pandemic-related research. For our potential for publications to be accounted for more than once. analysis, preprint articles related to COVID-19 biomedical Finally, for each publication, the geographic location of the first research were identified using two popular preprint servers: author's institution was recorded. medRxiv and bioRxiv. Specific search terms (see Multimedia Appendix 1) were used to identify and extract COVID-19 Analysis of Social Media, Preprint, and Publication preprint articles for each half-month period across the 6-month Activity study period for comparison with Twitter data. Furthermore, a The primary outcome of the analysis was to identify the peak follow-up review of the preprint articles pertaining to COVID-19 activity across the 3 platforms, as this shows how efficiently and gastroenterology that ultimately resulted in formal one platform could disseminate information as compared to the peer-reviewed journal publications was performed for the month others. ªActivityº was defined by the number of tweets (via of July 2020. This was done to account for the delay associated Twitter), or publications (when referring to preprint or with the formal publication of a preprint article. publication databases, such as bioRxiv and medRxiv or PubMed, respectively) produced. ªDisseminationº could be defined in For the analysis of peer-reviewed publications, the different ways as referenced by a previous systemic review PubMed±NCBI database was used to search for all publications report published by the Agency for Healthcare Research and pertaining to COVID-19 over the 6-month study period. The Quality (AHRQ); however, for the purpose of this study, it was

https://www.jmir.org/2021/6/e26956 J Med Internet Res 2021 | vol. 23 | iss. 6 | e26956 | p. 3 (page number not for citation purposes) XSL·FO RenderX JOURNAL OF MEDICAL INTERNET RESEARCH Taneja et al

defined as ªthe active and targeted distribution of information performed using STATA 15 (StataCorp, LLC) or interventions via determined channels using planned strategies software. Statistical significance was set at P<.05. All authors to a specific public health or clinical practice audienceº [20]. had access to the study data and reviewed and approved the In this study, the information distributed is the content of tweets final manuscript. or articles published in preprint repositories or accessible via the PubMed database. Of note, Twitter impressions is a Results convenient way to measure the exact distribution of the tweet content, as it calculates how many users would have been sent COVID-19±Related Publications and Twitter Activity a particular tweet based on the number of followers the user Trends who posted the tweet originally had. Secondary outcomes Over the 6-month study period from December 1, 2019, through included (1) peak activity in each platform overall and then by May 31, 2020, 73,079 tweets were identified from a total of GI subtopic, (2) peak activity in each platform by geographic 44,609 users, generating 207,039,610 impressions on the topic location, and (3) comparison of trends between the different of COVID-19. During this same period, 7164 publications platforms overall, as well as by GI subtopic (see Multimedia pertaining to COVID-19 were found to be indexed in PubMed Appendix 1). along with 4702 preprints archived in medRxiv and bioRxiv Summary statistics of baseline data for tweets, impressions, repositories. The overall summary of Twitter and research preprints and PubMed publications are presented as frequencies publication activity by half-month time interval is shown in for categorical data unless otherwise specified. Spearman's Table 1. Twitter activity, with regard to original tweets on the rank-order correlation was performed to determine the topic of COVID-19 did not appear until the latter half of January relationship between Twitter activity (ie, tweets and 2020, which resulted in 245 original tweets. This activity impressions) and PubMed publications overall, by organ system progressively increased thereafter and peaked during March and geographic location, as well as by Twitter activity and 16-31, 2020, with 20,660 original tweets before gradually preprint articles overall and by organ system. Analysis was decreasing over the remaining study interval (Figure 2).

Table 1. A comparison of productivity trends by half-month intervals of COVID-19±related publication, preprint, and Twitter activity from December 2019 through May 2020. Informational December January 2020 February 2020 March 2020 April 2020 May 2020 sources 2019 1-15 16-31 1-15 16-31 1-15 16-29 1-15 16-31 1-15 16-30 1-15 16-31

Publicationsa, n 0 0 0 34 135 180 342 588 1196 1586 1561 1541

Preprintsb, n 0 0 0 35 71 181 217 506 633 922 1051 1086

Tweetsc, n 0 0 0 245 592 815 13797 20660 13845 9636 7399 6090 Impressions 0 0 0 1,439,197 1,809,224 180,594 30,061,305 44,640,303 33,351,337 29,411,077 22,536,326 41,900,257 Impressions per 0 0 0 5874 3056 2320 2179 2179 2409 3052 3046 6880 tweet Unique users 0 0 0 165 362 552 9727 13034 8210 5647 3759 3153 Tweets per user 0 0 0 1.48 1.64 1.48 1.42 1.59 1.69 1.71 1.97 1.93

aPublications indexed in PubMed±NCBI database. bPreprints located in the medRxiv and bioRxiv repositories. cTweets and associated variables extracted using Symplur Signals search engine.

https://www.jmir.org/2021/6/e26956 J Med Internet Res 2021 | vol. 23 | iss. 6 | e26956 | p. 4 (page number not for citation purposes) XSL·FO RenderX JOURNAL OF MEDICAL INTERNET RESEARCH Taneja et al

Figure 2. Trend of COVID-19±related tweets, ratio of impressions to tweets, publications and preprints. Twitter activity from December 1, 2019, through May 31, 2020, captured at half-month intervals by using the Twitter analytics platform Symplur Signals. Preprints from medRxiv and bioRxiv repositories were also abstracted during this time period along with publications indexed in the PubMed-NCBI database.

A similar pattern of activity was observed among the number observed with regard to the number of tweets and Twitter users of Twitter users who posted tweets related to COVID-19, which (Figure 3). On average, the number of tweets per Twitter user increased from 165 users to 13,034 users between January and ranged from 1.48 to 1.97. Impressions generated per tweet were the latter half of March 2020. Twitter impressions followed a initially high (5874 impressions/tweet) in the latter half of similar pattern, with a peak observed during the second half of January 2020 but did not peak until the latter half of May 2020 March 2020. Interestingly, a second peak in impressions was (6880 impressions/tweet). Temporal trends are further detailed apparent during the latter half of May 2020, which was not in Figure 2.

Figure 3. Trend of COVID-19 Twitter impressions. Number of impressions generated from Twitter from December 1, 2019, through May 31, 2020, captured at half-month intervals by using the Twitter analytics platform Symplur Signals.

Scientific COVID-19±related articles indexed in PubMed as similar trajectory as Twitter activity, with both the first well as preprints in medRxiv and bioRxiv servers followed a peer-reviewed articles and the first preprints becoming available

https://www.jmir.org/2021/6/e26956 J Med Internet Res 2021 | vol. 23 | iss. 6 | e26956 | p. 5 (page number not for citation purposes) XSL·FO RenderX JOURNAL OF MEDICAL INTERNET RESEARCH Taneja et al

during the second half of January 2020 [21]. However, unlike publications (ρ=0.58), as well as between Twitter activity and Twitter activity that peaked in the second half of March 2020, number of preprints (ρ=0.57) across the study duration (P<.001 publications and preprints reached peak activity around the for both; Table 2). second half of April 2020. Notably, we observed a parallel rise Similarly, there was a moderately strong association between in the number of preprints and PubMed publications (Table 1 the number of Twitter impressions and PubMed publications and Figure 2). (ρ=0.56, P<.001), as well as between the number of Twitter A moderately strong correlation was demonstrated between impressions and preprints (ρ=0.54, P<.001; Table 3). Twitter activity (ie, number of tweets) and number of PubMed

Table 2. Correlation between Twitter activity (tweets) and PubMed publications and between Twitter activity (tweets) and preprints by organ system. Italicized values indicate statistical significance.

Organ system Tweets and PubMed publications (ρ) P value Tweets and preprints (ρ) P value Overall trend 0.58 <.001 0.57 <.001 Trend by organ system Pulmonology or critical care 0.8 .002 0.8 .003 Luminal gastroenterology 0.7 .009 0.6 .03 Hepatology 0.7 .006 0.7 .009 Inflammatory bowel disease 0.7 .006 0.5 .07 Pancreatology 0.4 .20 0.4 .30 Gastrointestinal endoscopy 0.7 .007 0.7 .02

Table 3. Correlation between Twitter activity (impressions) and PubMed publications and between Twitter activity (impressions) and preprints by organ system. Italicized values indicate statistical significance.

Organ system Impressions and PubMed publications (ρ) P value Impressions and preprints (ρ) P value Overall trend 0.56 <.001 0.54 <.001 Trend by organ system Pulmonology or critical care 0.8 .001 0.8 .002 Luminal gastroenterology 0.7 .009 0.7 .006 Hepatology 0.7 .005 0.7 .006 Inflammatory bowel disease 0.8 .004 0.4 .20 Pancreatology 0.5 .07 0.3 .30 Gastrointestinal endoscopy 0.8 .004 0.7 .02

within the field, the majority of tweets were on the topics of COVID-19±Related Twitter, Publication, and Preprint pancreatology (5804/73,079, 7.9%), followed by luminal Content Classified by Organ System Topic gastroenterology (3318/73,079, 4.5%), inflammatory bowel Overview disease (2818/73,079, 3.9%), GI endoscopy (1764/73,079, 2.4%), and hepatology (583/73,079, 0.8%). With regard to Analysis of Twitter, publication, and preprint data pertaining gastroenterology-related publications, the majority of articles to the effects of COVID-19 on specific organ system topics are were on the topic of hepatology (236/7250, 3.3%) followed by outlined in Table S1 of Multimedia Appendix 2, and trends are GI endoscopy (111/7250, 1.5%), luminal gastroenterology illustrated in Figure 4 and Figure 5. The majority of (64/7250, 0.9%), IBD (30/7250, 0.4%), and pancreatology COVID-19±related tweets (58,792/73,079, 80.4%), publications (10/7250, 0.1%). Preprint publications were primarily on the (6713/7164, 93.7%), and preprint articles (4567/4702, 97.1%) topics of luminal gastroenterology (21/4702, 0.4%), hepatology covered the topic of pulmonology or critical care. (18/4702, 0.4%), and IBD (3/4702, 0.1%) (see Table S1 in Gastroenterology was a small subset of these topics; however, Multimedia Appendix 2).

https://www.jmir.org/2021/6/e26956 J Med Internet Res 2021 | vol. 23 | iss. 6 | e26956 | p. 6 (page number not for citation purposes) XSL·FO RenderX JOURNAL OF MEDICAL INTERNET RESEARCH Taneja et al

Figure 4. Trend of Twitter, publication, and preprint activity related to COVID-19 and pulmonary or critical care topics. Comparison of the number of tweets posted with articles published in peer-reviewed journals and preprints published pertaining to COVID 19 and pulmonary or critical care at half-month intervals between December 2019 and May 2020.

https://www.jmir.org/2021/6/e26956 J Med Internet Res 2021 | vol. 23 | iss. 6 | e26956 | p. 7 (page number not for citation purposes) XSL·FO RenderX JOURNAL OF MEDICAL INTERNET RESEARCH Taneja et al

Figure 5. Trend of COVID-19 and gastroenterology subspecialty±related Twitter, publication, and preprint activity. Trend in the number of (A) tweets, (B) publications, and (C) preprints published pertaining to COVID 19 and gastroenterology subspecialty topics at half-month intervals between December 2019 and May 2020. GI: gastrointestinal; IBD: inflammatory bowel disease.

was observed in the latter half of March (Figure 4 and Table S1 Pulmonology and Critical Care in Multimedia Appendix 2). Approximately 59,000 COVID-19±related tweets analyzed across the study period were on the topic of pulmonology or A total of 6713 peer-reviewed articles on the topic of critical care. The most significant increase in tweets on this pulmonology or critical care and COVID-19 were indexed in topic occurred between February and March 2020, with a nearly PubMed during the 6-month study period, and these first 22-fold increase, and an ultimate peak in activity (n=16,489) appeared in January 2020. The most significant increase (ie, 2-fold) in the number of publications on COVID-19 and

https://www.jmir.org/2021/6/e26956 J Med Internet Res 2021 | vol. 23 | iss. 6 | e26956 | p. 8 (page number not for citation purposes) XSL·FO RenderX JOURNAL OF MEDICAL INTERNET RESEARCH Taneja et al

pulmonology or critical care was observed between the latter the second half of March and early April 2020, with a 5-fold half of March 2020 and the start of April 2020. As compared increase as detailed in Table S1 in Multimedia Appendix 2. to peer-reviewed publications, there were approximately A total of 45 COVID-19 and gastroenterology±related preprints one-third fewer preprint articles (n=4567) in medRxiv and were archived in medRxiv and bioRxiv servers over the study bioRxiv related to COVID-19 and pulmonology or critical care period. Longitudinal analysis showed that the number of identified during the study period. Moreover, on the topic of preprints on the topic of gastroenterology peaked in the latter pulmonology or critical care and COVID-19, the longitudinal half of April 2020. When further stratified by subspecialty, trend in preprint article availability appears to have paralleled unlike that observed with peer-reviewed publications, the publications indexed in PubMed; however, for preprint articles, majority of preprints covered luminal gastroenterology (21/45, the most significant rise was observed 2 weeks prior to that 46.7%), followed by hepatology (18/45, 40%), IBD (3/45, observed with PubMed publications, specifically between the 6.7%), GI endoscopy (2/45, 4.4%), and pancreatology (1/45, first and second half of March (Figure 4 and Table S1 in 2.2%) (Table S1 in Multimedia Appendix 2). Multimedia Appendix 2). Similar to pulmonology or critical care±related content, there There was a strong correlation between both the number of was a strong correlation between tweets and peer-reviewed COVID-19±related tweets and peer-reviewed publications publications (ρ=0.6, P=.03) as well as between tweets and (ρ=0.8, P=.002) as well as between the number of preprints (ρ=0.7, P=.009) on the topic of luminal COVID-19±related tweets and preprints (ρ=0.8, P=.003, gastroenterology. Additionally, a strong correlation was respectively; Table 2) on the topic of pulmonology or critical identified between both the number of tweets and peer-reviewed care. Similarly, there was a strong correlation between both publications (ρ=0.7, P=.006) as well as between tweets and pulmonology or critical care±related Twitter impressions and preprints (ρ=0.7, P=.009) on the topic of COVID-19 and publications (ρ=0.8, P=.001) as well as between Twitter hepatology. A similarly strong correlation was observed between impressions and preprints (ρ=0.8, P=.002; Table 3). the number of tweets and PubMed publications on the topic of Gastroenterology GI endoscopy (ρ=0.7, P=.007), the number of tweets and A total of 14,285 tweets concerning the field of gastroenterology preprints on the topic of GI endoscopy (ρ=0.7, P=.02), and the and COVID-19 (encompassing subspecialty fields of luminal number of tweets and peer-reviewed publications on the topic gastroenterology, IBD, hepatology, GI endoscopy, and of COVID-19 and IBD (ρ=0.7, P=.008). In contrast, no pancreatology) were identified during the study period (Figure significant correlation was identified between tweets and 5 and Table S1 in Multimedia Appendix 2). Among all tweets peer-reviewed publications or preprints on the topic of recorded during the 6-month study period, 19.6% COVID-19 and pancreatology (Table 2). (14,287/73,079) were on the topic of COVID-19 and Regarding COVID-19 and luminal gastroenterology content, a gastroenterology. The longitudinal trend in number of strong correlation was observed between both Twitter gastroenterology-related tweets (including subspecialty impressions and peer-reviewed publications (ρ=0.7, P=.009) gastroenterology fields) paralleled that observed with as well as between Twitter impressions and preprints (ρ=0.7, pulmonology or critical care±related tweets, with an approximate P=.006). Similarly, strong correlations were identified between 45-fold increase in the number of tweets spanning the latter half Twitter impressions and peer-reviewed publications on the of February 2020 (n=75) and peaking in the second half of topics of COVID-19 and hepatology (ρ=0.7, P=.005), IBD March 2020 (n=4171). When further stratified by subspecialty (ρ=0.8, P=.004), and GI endoscopy (ρ=0.8, P=.004). There was field, the majority of COVID-19 and gastroenterology±related no significant correlation between Twitter impressions and tweets were on the topic of pancreatology (5804/14,287, 40.6%), publications on the topic of COVID-19 and pancreatology. In followed by luminal gastroenterology (3318/14,287, 23.2%), evaluating the association between Twitter impressions and IBD (2647/14,287, 19.7%), GI endoscopy (1764/14,287, 12.3%), preprints, strong associations were found on the topics of and hepatology (583/14,287, 4.1%). hepatology (ρ=0.7, P=.006) and GI endoscopy (ρ=0.7, P=.02), A total of 449 peer-reviewed publications related to COVID-19 whereas no associations were found pertaining to the topics of and gastroenterology were identified in PubMed during the IBD and pancreatology (Table 3). study period. In contrast to Twitter activity, the majority of COVID-19 Twitter and Publication Content by these publications were on the topic of hepatology (235/449, Geographic Location 52.3%) followed by GI endoscopy (111/449, 24.6%), luminal gastroenterology (64/449, 14.2%), IBD (30/449, 6.7%), and The top 5 countries with the highest number of pancreatology (10/449, 2.2%). Similar to Twitter activity, COVID-19±related tweets posted globally over the 6-month PubMed publications on the topics of luminal gastroenterology study period included the United States (7331/22,215, 33.0%), and hepatology first appeared in the latter half of January 2020. followed by the United Kingdom (4229/22,215, 19.3%), Spain The most significant increase in COVID-19 and liver±related (1527/22,215, 6.8%), Canada (1174/22,215, 5.3%), and publications was observed between the latter half of March Australia (673/22,215, 3.0%). China generated the highest (n=12) and early April 2020 (n=31), with an over 2.5-fold number of peer-reviewed publications indexed in PubMed increase in the number of publications on this topic. Luminal (1768/6352, 27.8%) throughout the study period, followed by gastroenterology±related publications, which first appeared in Italy (915/6352, 14.4%), the United States (389/6352, 6.1%), the latter half of February 2020, significantly increased between France (348/6352, 5.5%), and India (303/6352, 4.8%). Figure 6 illustrates the countries with the highest Twitter activity and

https://www.jmir.org/2021/6/e26956 J Med Internet Res 2021 | vol. 23 | iss. 6 | e26956 | p. 9 (page number not for citation purposes) XSL·FO RenderX JOURNAL OF MEDICAL INTERNET RESEARCH Taneja et al

peer-reviewed publication activity. The top 20 countries with correlation between the number of tweets and peer-reviewed the highest number of tweets and the highest number of publications in both the United States (ρ=0.8, P=.005) and the peer-reviewed publications are listed in Tables S3 and S4, United Kingdom (ρ=0.8, P=.01; see Table S2 in in Multimedia respectively, in Multimedia Appendix 2. There was a strong Appendix 2).

Figure 6. Heat maps illustrating the total number of COVID-19±related tweets (A) across the globe and (C) in the United States, as well as publications indexed in the PubMed-NCBI represented (B) across the globe and (D) in the United States, over the 6-month study period (December 2019 through May 2020). Numbers are represented on the spectrum from the least (yellow) to the highest amount (maroon), as detailed in the legend accompanying each map. Countries or US states shaded white indicate the absence of data for those regions.

Within the United States, when analyzing both Twitter activity and peer-reviewed publications, the state of New York had the Discussion highest COVID-19±related activity (11% of tweets and 39.1% Principal Findings of publications) during the study period, followed closely by California (10.3% of tweets and 36.6% of publications). Figure Since its emergence in December 2019, the novel coronavirus 6 illustrates the Twitter activity and publication activity in the SARS-CoV-2 has triggered an unparalleled global response in US states. The top 20 US states with the highest number of the fields of science, medicine, public health, and technology. tweets and peer-reviewed publications are listed in Tables S5 Considering its highly contagious nature, along with the paucity and S6, respectively, in Multimedia Appendix 2. of knowledge and current lack of effective treatment modalities to combat the infection, the need for rapid sharing and COVID-19 Twitter Content by User Stakeholder dissemination of information has been paramount. In this study, Designation or Category we assessed the dissemination of COVID-19±related information Twitter user data entered as the Twitter user's self-designated via preprint services, formal peer-reviewed publications, and health care stakeholder role was analyzed. For the topics of through the global reach of the social media platform Twitter. pulmonology or critical care, luminal gastroenterology, Specifically, we observed that during the second half of March hepatology, IBD, and GI endoscopy, the top 2 most active 2020, when COVID-19 was continuing to spread rapidly stakeholder categories were doctors/physicians and prompting various nations, including the United States, to enter researchers/academic users. Furthermore, for the topic of a state of lockdown, social media activity on Twitter was at its pancreatology, advocacy organizations and patient advocates peak, with almost 7000 impressions per tweet analyzed. were the most active stakeholder users. The top 15 most active Furthermore, we observed that Twitter activity was strongly users and their stakeholder roles categorized by Twitter activity correlated with the published scientific data available to the for each organ system topic are further detailed in Table S7 in general public. Although COVID-19 has been predominantly Multimedia Appendix 2. linked with severe pulmonary complications, approximately 20% of the conversations on social media was related to the

https://www.jmir.org/2021/6/e26956 J Med Internet Res 2021 | vol. 23 | iss. 6 | e26956 | p. 10 (page number not for citation purposes) XSL·FO RenderX JOURNAL OF MEDICAL INTERNET RESEARCH Taneja et al

field of gastroenterology with specific discussions related to COVID-19 have been less commonly reported in the literature hepatology, GI endoscopy, luminal gastroenterology, IBD, and as compared to other GI symptoms, we unexpectedly observed pancreatology. Social media activity was strongly associated that Twitter activity on the topic of pancreatology was the with the availability of published data pertaining to all highest among all gastroenterology subtopics. However, there gastroenterology topics with the exception of pancreatology. was no statistically significant correlation between Twitter and Finally, in our analysis of data by geographic region of publication activity or between Twitter and preprint activity on publication, social media activity was most prominent in the the topic of pancreatology, thus implying that the rise of social regions most affected by the pandemic both globally (in regions media activity did not necessarily reflect the same rise in activity where social media via Twitter is not banned) and within the in bioRxiv, medRxiv, or PubMed activity. This could possibly United States, with strong associations between social media be explained by reviewing the stakeholder data for and publication data. pancreatology-related tweets. For each of the other gastroenterology subfields, stakeholder demographics fell In the longitudinal assessment of publication activity, the authors largely under the categories of physicians, nonmedical doctors, observed that peak social media activity predated peak PubMed or researchers and academic users; however, for the topic of publications by approximately 30 days. One possible explanation ªpancreas,º the most active stakeholders appeared to be for this lag interval is the technical delay that typically occurs advocacy organizations or patient advocates. Upon further after acceptance of a peer-reviewed manuscript by a journal and review of the top 50 associated hashtags used for the ªpancreasº prior to indexing in PubMed. It is worth highlighting that around topic, those related to cystic fibrosis support and awareness the peak of PubMed publication activity, we also witnessed a groups were the most common. Patient advocacy organizations parallel rise in the preprint repository activity. Historically, the not only play an important part in espousing awareness of use of preprint repositories has allowed researchers to ªclaim specific diseases and patient populations, but they also serve the spaceº or even to ªpublish firstº in contentious fields of an essential role in the dissemination of information on their science and research. However, in the face of an evolving behalf, typically in the form or raising awareness on the internet, pandemic, preprint repositories are serving as a new mode of lobbying directly for change within the government of other scientific communication, bypassing the typical lengthy institutions, and via marketing and outreach, but likely only peer-review process and thus allowing for faster dissemination indirectly in the context of research activity. Therefore, this and communication of research and clinical findings related to increase in Twitter activity may not translate to a proportionate COVID-19. In fact, in March 2020, when the World Health increase in research publication activity on the topic of Organization officially declared COVID-19 a pandemic, 8830 COVID-19 and pancreatology as demonstrated in this study. biomedical preprints were published, a 142% increase from the Using this topic as an example, further investigation diving previous year, and medRxiv page views had increased to 15 deeper into the content of tweets, specific users, and how they million a month, as compared to 1 million a month prior to the translate directly into published work would be helpful to start of the pandemic [22]. As of July 1, 2020, we were able to document the direct impact social media has on research activity. document that nearly half of all preprint articles related to COVID-19 and gastroenterology have subsequently made their Limitations way into reputable scientific journals, further supporting this There are several limitations to this study that are worth noting. theory. The first major limitation is that social media platforms, such Notably, we observed that physicians, nonmedical doctors, and as Twitter, are not available in several countries, including scientific researchers constituted the lead stakeholder activity China. This could help to explain China's lead in research for social media use overall during the COVID-19 pandemic. publication activity as compared with other regions, as This observation suggests that even during the early stages of publication is likely one of the primary modalities used in this the COVID-19 pandemic, social media became an increasingly country for disseminating information. Other social media sought-after tool, likely for the purpose of communicating platforms, including WeChat and Sina Weibo, are used in China; medical and scientific information. This study confirmsÐby however, information regarding the use of these other social documenting health care stakeholder activityÐthat the scientific media platforms within China is limited to date. Future studies and medical communities leveraged social media platforms are needed to assess social media activity on these alternative during the early stages of the COVID-19 pandemic more so platforms and their association with publications as well as how than other health care stakeholders such as patients or advocacy they compared to the use of Twitter in other nations, such as groups. Previous studies performed during nonpandemic times the United States (where Twitter is not banned). Second, have demonstrated conflicting evidence regarding the impact although we were able to account for and reassign duplicate of social media on scientific publications related to citation publications for the various categorizations performed, a similar impact, metrics, and viewability [23-26]. This study clearly approach could not be guaranteed for Twitter data. The very illustrated that social media activity was greater and also peaked nature of tweets allows for other users to publish an original earlier than any of the publication modalities, which is important tweet to their account generating additional impressions as it is to consider when investigating how social media could sent to their followers (also known as a ªretweetº). Therefore, potentially impact research and scientific work in future. limiting duplicate tweets may artificially decrease activity. More importantly, certain tweets, or even retweets, may have been The gastroenterology subtopic of pancreatology is a good assigned to more than one topic area owing to limitations of the example of this potential influence, or rather, the lack thereof. Symplur software. Hence, we were unable to limit tweets to a Although pancreatic manifestations (eg, acute pancreatitis) of

https://www.jmir.org/2021/6/e26956 J Med Internet Res 2021 | vol. 23 | iss. 6 | e26956 | p. 11 (page number not for citation purposes) XSL·FO RenderX JOURNAL OF MEDICAL INTERNET RESEARCH Taneja et al

single topic, and as such, this may have artificially boosted the the GI community. social media tools like Twitter can be an overall number in certain organ system topics, thereby effective method for educating and informing audiences in real potentially skewing the results of this study. time and via an interactive approach, a feat that cannot always be achieved with more conventional methods (ie, scientific Conclusions publications). The new media age has resulted in a number of In conclusion, this study demonstrates patterns in the utility of novel avenues for the distribution of information, including social media and publicationsÐboth in preprint repositories as Twitter. Utilizing a single modality for dissemination of health well as peer-reviewed journalsÐfor the rapid dissemination of care discoveries or information has been shown by the AHRQ information during a global pandemic of infectious disease. As to not be as effective as utilizing multiple modalities. It may the world faces this unprecedented public health emergency, therefore be time for the medical and scientific communities to this study has reflected on shifting worldwide trends from solely cultivate formal social media platforms as effective tools for traditional methods of disseminating information (ie, via data sharing and collaboration to augment existing modalities publications) to more contemporary methods, specifically among of archival publication.

Conflicts of Interest None declared.

Multimedia Appendix 1 Search terms used for Twitter, PubMed, and preprint repositories. [DOCX File , 18 KB-Multimedia Appendix 1]

Multimedia Appendix 2 Supplementary tables summarizing additional information pertaining to COVID-19±related Twitter, publication, and preprint activity by organ system per half-month time intervals. [DOCX File , 40 KB-Multimedia Appendix 2]

References 1. Coronavirus Disease (COVID-19)Dashboard. World Health Organization. URL: https://covid19.who.int/ [accessed 2021-03-03] 2. Zhang J, Dong X, Cao Y, Yuan Y, Yang Y, Yan Y, et al. Clinical characteristics of 140 patients infected with SARS-CoV-2 in Wuhan, China. Allergy 2020 Jul;75(7):1730-1741. [doi: 10.1111/all.14238] [Medline: 32077115] 3. Mao , Liang J, Shen J, Ghosh S, Zhu L, Yang H, Chinese Society of IBD¸ Chinese Elite IBD Union, Chinese IBD Quality Care Evaluation Center Committee. Implications of COVID-19 for patients with pre-existing digestive diseases. Lancet Gastroenterol Hepatol 2020 May;5(5):425-427 [FREE Full text] [doi: 10.1016/S2468-1253(20)30076-5] [Medline: 32171057] 4. Ding S, Liang TJ. Is SARS-CoV-2 also an enteric pathogen with potential fecal-oral transmission? a COVID-19 virological and clinical review. Gastroenterology 2020 Jul;159(1):53-61 [FREE Full text] [doi: 10.1053/j.gastro.2020.04.052] [Medline: 32353371] 5. Fan Z, Chen L, Li J, Cheng X, Yang J, Tian C, et al. Clinical features of COVID-19-related liver functional abnormality. Clin Gastroenterol Hepatol 2020 Jun;18(7):1561-1566 [FREE Full text] [doi: 10.1016/j.cgh.2020.04.002] [Medline: 32283325] 6. Donato MF, Invernizzi F, Lampertico P, Rossi G. Health status of patients who underwent liver transplantation during the coronavirus outbreak at a large center in Milan, Italy. Clin Gastroenterol Hepatol 2020 Aug;18(9):2131-2133.e1 [FREE Full text] [doi: 10.1016/j.cgh.2020.04.041] [Medline: 32334081] 7. D©Amico F, Peyrin-Biroulet L, Danese S. Inflammatory bowel diseases and COVID-19: the invisible enemy. Gastroenterology 2020 Jun;158(8):2302-2304 [FREE Full text] [doi: 10.1053/j.gastro.2020.04.032] [Medline: 32305331] 8. Rubin DT, Abreu MT, Rai V, Siegel CA, International Organization for the Study of Inflammatory Bowel Disease. Management of patients with Crohn©s disease and ulcerative colitis during the coronavirus disease-2019 pandemic: results of an international meeting. Gastroenterology 2020 Jul;159(1):6-13.e6 [FREE Full text] [doi: 10.1053/j.gastro.2020.04.002] [Medline: 32272113] 9. Rodríguez-Lago I, Ramírez de la Piscina P, Elorza A, Merino O, Ortiz de Zárate J, Cabriada JL. Characteristics and prognosis of patients with inflammatory bowel disease during the SARS-CoV-2 pandemic in the Basque Country (Spain). Gastroenterology 2020 Aug;159(2):781-783 [FREE Full text] [doi: 10.1053/j.gastro.2020.04.043] [Medline: 32330477] 10. Repici A, Maselli R, Colombo M, Gabbiadini R, Spadaccini M, Anderloni A, et al. Coronavirus (COVID-19) outbreak: what the department of endoscopy should know. Gastrointest Endosc 2020 Jul;92(1):192-197 [FREE Full text] [doi: 10.1016/j.gie.2020.03.019] [Medline: 32179106]

https://www.jmir.org/2021/6/e26956 J Med Internet Res 2021 | vol. 23 | iss. 6 | e26956 | p. 12 (page number not for citation purposes) XSL·FO RenderX JOURNAL OF MEDICAL INTERNET RESEARCH Taneja et al

11. Cennamo V, Bassi M, Landi S, Apolito P, Ghersi S, Dabizzi E, et al. Redesign of a GI endoscopy unit during the COVID-19 emergency: a practical model. Dig Liver Dis 2020 Oct;52(10):1178-1187 [FREE Full text] [doi: 10.1016/j.dld.2020.05.007] [Medline: 32425734] 12. Dunford D, Dale B, Stylianou N, Lowther E, Ahmed M, de la Torre Arenas I. Coronavirus: The world in lockdown in maps and charts. BBC News. 2020 Apr 7. URL: https://www.bbc.com/news/world-52103747 [accessed 2021-04-12] 13. Depoux A, Martin S, Karafillakis E, Preet R, Wilder-Smith A, Larson H. The pandemic of social media panic travels faster than the COVID-19 outbreak. J Travel Med 2020 May 18;27(3):taaa031 [FREE Full text] [doi: 10.1093/jtm/taaa031] [Medline: 32125413] 14. Merchant RM, Lurie N. Social media and emergency preparedness in response to novel coronavirus. JAMA 2020 May 26;323(20):2011-2012. [doi: 10.1001/.2020.4469] [Medline: 32202611] 15. Pershad Y, Hangge PT, Albadawi H, Oklu R. Social medicine: Twitter in healthcare. J Clin Med 2018 May 28;7(6):121. [doi: 10.3390/jcm7060121] [Medline: 29843360] 16. Merchant RM. Evaluating the potential role of social media in preventive health care. JAMA 2020 Feb 04;323(5):411-412. [doi: 10.1001/jama.2019.21084] [Medline: 31922532] 17. Sinnenberg L, Buttenheim AM, Padrez K, Mancheno C, Ungar L, Merchant RM. Twitter as a tool for health research: a . Am J Public Health 2017 Jan;107(1):e1-e8. [doi: 10.2105/AJPH.2016.303512] [Medline: 27854532] 18. Brady RRW, Chapman SJ, Atallah S, Chand M, Mayol J, Lacy AM, et al. #colorectalsurgery. Br J Surg 2017 Oct;104(11):1470-1476. [doi: 10.1002/bjs.10615] [Medline: 28881004] 19. Symplur Signals. URL: https://www.symplur.com/products/signals/ [accessed 2021-05-24] 20. McCormack L, Sheridan S, Lewis M, Boudewyns V, Melvin CL, Kistler C, et al. Communication and dissemination strategies to facilitate the use of health-related evidence. Evid Rep Technol Assess (Full Rep) 2013 Nov(213):1-520. [doi: 10.23970/ahrqepcerta213] [Medline: 24423078] 21. Cohen J, Normile D. New SARS-like virus in China triggers alarm. Science 2020 Jan 17;367(6475):234-235. [doi: 10.1126/science.367.6475.234] [Medline: 31949058] 22. Teixeira da Silva JA, Tsigaris P, Erfanmanesh M. Publishing volumes in major databases related to Covid-19. 2020 Aug 28:1-12 [FREE Full text] [doi: 10.1007/s11192-020-03675-3] [Medline: 32904414] 23. Fox CS, Bonaca MA, Ryan JJ, Massaro JM, Barry K, Loscalzo J. A randomized trial of social media from Circulation. Circulation 2015 Jan 06;131(1):28-33. [doi: 10.1161/CIRCULATIONAHA.114.013509] [Medline: 25406308] 24. Fox CS, Gurary EB, Ryan J, Bonaca M, Barry K, Loscalzo J, et al. Randomized controlled trial of social media: effect of increased intensity of the intervention. J Am Heart Assoc 2016 Apr 27;5(5):e003088 [FREE Full text] [doi: 10.1161/JAHA.115.003088] [Medline: 27121850] 25. Eysenbach G. Can tweets predict citations? metrics of social impact based on Twitter and correlation with traditional metrics of scientific impact. J Med Internet Res 2011 Dec 19;13(4):e123 [FREE Full text] [doi: 10.2196/jmir.2012] [Medline: 22173204] 26. Bardus M, El Rassi R, Chahrour M, Akl EW, Raslan AS, Meho LI, et al. The use of social media to increase the impact of health research: systematic review. J Med Internet Res 2020 Jul 06;22(7):e15607 [FREE Full text] [doi: 10.2196/15607] [Medline: 32628113]

Abbreviations ACE2: angiotensin-2 AHRQ: Agency for Healthcare Research and Quality GI: gastrointestinal IBD: inflammatory bowel disease MERS: Middle East Respiratory Syndrome SARS: severe acute respiratory syndrome coronavirus

Edited by C Basch; submitted 05.01.21; peer-reviewed by C Hudak, S Mathews, K Reuter; comments to author 11.02.21; revised version received 05.05.21; accepted 17.05.21; published 17.06.21 Please cite as: Taneja SL, Passi M, Bhattacharya S, Schueler SA, Gurram S, Koh C Social Media and Research Publication Activity During Early Stages of the COVID-19 Pandemic: Longitudinal Trend Analysis J Med Internet Res 2021;23(6):e26956 URL: https://www.jmir.org/2021/6/e26956 doi: 10.2196/26956 PMID: 33974550

https://www.jmir.org/2021/6/e26956 J Med Internet Res 2021 | vol. 23 | iss. 6 | e26956 | p. 13 (page number not for citation purposes) XSL·FO RenderX JOURNAL OF MEDICAL INTERNET RESEARCH Taneja et al

©Sonia L Taneja, Monica Passi, Sumona Bhattacharya, Samuel A Schueler, Sandeep Gurram, Christopher Koh. Originally published in the Journal of Medical Internet Research (https://www.jmir.org), 17.06.2021. This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on https://www.jmir.org/, as well as this copyright and license information must be included.

https://www.jmir.org/2021/6/e26956 J Med Internet Res 2021 | vol. 23 | iss. 6 | e26956 | p. 14 (page number not for citation purposes) XSL·FO RenderX